A high degree of heart rate (HR) variability is found in persons with normal hearts, whereas low HR variability can be found in patients with severe coronary artery disease, congestive heart failure and diabetic neuropathy. Two weeks after acute myocardial infarction, low HR variability predicted reduced long-term survival even after adjusting for clinical risk indicators, left ventricular ejection fraction, HR and ventricular arrhythmias. The present study elucidated the causes of differences in HR and HR variability between patients with low and high HR variability. In a matched-pair study, 10 patients with low HR variability (24-hour standard deviation of N-N intervals less than 50 ms) were randomly selected. For each of these 10 patients, a control patient with high HR variability (24-hour standard deviation of N-N intervals greater than or equal to 100 ms), matched for age, left ventricular ejection fraction and rales in the coronary care unit was selected. Patients who were taking either digitalis or beta-adrenergic blocking drugs were excluded. Analysis of 24-hour electrocardiograms showed that for the low HR variability group compared with the high: (1) the daytime and nighttime average HR was faster; (2) the difference between daytime and nighttime HR was less; (3) the proportion of differences greater than 50 ms between successive N-N intervals was smaller; and (4) the number of HR "spikes" per day (increase in HR greater than or equal to 10 beats/min, lasting from 3 to 15 minutes) was less.(ABSTRACT TRUNCATED AT 250 WORDS)